Concordance heads (QK circuits) could serve as the consistency-checking circuit for distinguishing intended vs. unintended outputs

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Each attention head has two largely independent computations: a QK circuit computing the attention pattern and an OV circuit computing the effect if attended toclaim0.778
  Key decomposition enabling separate analysis of where attention goes and what it does
• The prefill detection task may involve concordance heads that measure the likelihood of the output given prior activationsclaim0.773
  Speculation that QK circuit 'concordance heads' underlie the ability to distinguish intended from unintended outputs.
• Circuits could act as an epistemic foundation for interpretability by breaking down model behavior into falsifiable statements about small subgraphs.claim0.746
  Normative vision for how the circuits agenda could resolve the pre-paradigmatic state of interpretability
• Concordance headsconcept0.745
  QK circuit heads hypothesized to measure likelihood of an output given prior activations, used in prefill detection.
• Options contracts and ambiguous centers both encode multiple simultaneous readings as counterfactual structures.claim0.740
• Precise specification informs use and implementation without entangling them.claim0.738
• Precision is the unified mechanism for attention and meta-awareness across hierarchical levels.claim0.730
  Core theoretical claim: attention = precision over sensory evidence; meta-awareness = precision over attentional states; meditation = precision modulation.
• The sensitivity to think/don't think instructions may be achieved via a circuit that tags tokens as attention-worthy based on instructions or incentiveshypothesis0.726
  Mechanism for how the model modulates representation strength.